Graph-Based-Text-Classification

This code is for our proposed model in DL-NLP course.

• Steps to run this model:

1. Datasets of MR, SST-2, R8 and 20ng should be put under Data/ and path needs to be updated in config.py file.
2. For parameter tuning, use config.py file and change parameters.
3. python3 train.py

Baselines:

• TF-IDF with Logistic Regression:TF-IDF + LR
• LSTM with pre-trained GloVe embeddings(d=300) : LSTM - GloVe
• Code for TGCN and VGCN-BERT: adjacency.ipynb + gcn.py + train.py

Dataset Statistics:

Results:

Project Report- Final_Project_Report.pdf